In a normal Virtual Reality (VR) display system, discomfort may arise if movement of the human's head and rotation of eyeballs do not match with an image entering human eyes. Thus, fast follow-up and response time are very important for VR display.
In e.g. a VR display environment, a nine-axis sensor and infrared eye-tracking camera is generally used to collect a scenes focused by the human eyes. When a point focused by the eyeballs changes, a display panel may only scan a small portion of the scene. In conventional solutions, when the displayed scene changes, a Central Processing Unit (CPU) notifies a Graphics Processing Unit (GPU) of rendering a new image, but display information of a new scene cannot be sent until the display panel has finished scanning the previous frame of screen, which causes that the response time becomes slower, or even has nearly one frame of time wasted, affecting objective experience; otherwise, there will be a tearing effect.
Therefore, how to reduce the response time upon change of the scene is an urgent problem to be solved.